Delft University of Technology

Exergy return on exergy investment analysis of natural-polymer (Guar-Arabic gum)enhanced oil recovery process

Hassan, Anas M.; Ayoub, M.; Eissa, M.; Musa, T.; Bruining, Hans; Farajzadeh, R.

DOI10.1016/j.energy.2019.05.137Publication date2019Document VersionFinal published versionPublished inEnergy

Citation (APA)Hassan, A. M., Ayoub, M., Eissa, M., Musa, T., Bruining, H., & Farajzadeh, R. (2019). Exergy return onexergy investment analysis of natural-polymer (Guar-Arabic gum) enhanced oil recovery process. Energy,181, 162-172. https://doi.org/10.1016/j.energy.2019.05.137

Important noteTo cite this publication, please use the final published version (if applicable).Please check the document version above.

CopyrightOther than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consentof the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Takedown policyPlease contact us and provide details if you believe this document breaches copyrights.We will remove access to the work immediately and investigate your claim.

This work is downloaded from Delft University of Technology.For technical reasons the number of authors shown on this cover page is limited to a maximum of 10.

Green Open Access added to TU Delft Institutional Repository

‘You share, we take care!’ – Taverne project

https://www.openaccess.nl/en/you-share-we-take-care

Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Journal of Industrial and Engineering Chemistry 72 (2019) 133–143

A 2-D simulation study on CO2 soluble surfactant for foam enhanced oilrecovery

Yongchao Zenga, Rouhi Farajzadehb,c,*, Sibani L. Biswala, George J. Hirasakia

aRice University, 6100 Main St., MS-362, Department of Chemical and Biomolecular Engineering, Houston, TX, 77005 USAb Shell Global Solutions International, 2288GS Rijswijk, The NetherlandscDelft University of Technology, Delft, 2628CN, The Netherlands

A R T I C L E I N F O

Article history:Received 3 October 2018Received in revised form 25 November 2018Accepted 6 December 2018Available online 14 December 2018

Keywords:Nonionic surfactantPartition coefficientCO2

FoamGas breakthroughMobility controlEnhanced oil recovery (EOR)Foam simulation

A B S T R A C T

This paper probes the transport of CO2 soluble surfactant for foaming in porous media. We numericallyinvestigate the effect of surfactant partitioning between the aqueous phase and the gaseous phase onfoam transport for subsurface applications when the surfactant is injected in the CO2 phase. A 2-Dreservoir simulation is developed to quantify the effect of surfactant partition coefficient on thedisplacement conformance and CO2 sweep efficiency. A texture-implicit local-equilibrium foam model isembedded to describe how the partitioning of surfactant between water and CO2 affects the CO2 foammobility control when surfactant is injected in the CO2 phase. We conclude that when surfactant hasapproximately equal affinity to both the CO2 and the water, the transport of surfactant is in line with thegas propagation and therefore the sweep efficiency is maximized. Too high affinity to water (smallpartition coefficient) results in surfactant retardation whereas too high affinity to CO2 (large partitioncoefficient) leads to weak foam and insufficient mobility reduction. This work sheds light upon the designof water-alternating-gas-plus-surfactant-in-gas (WAG + S) process to improve the conventional foamprocess with surfactant-alternating-gas (SAG) injection mode during which significant amount ofsurfactant could possibly drain down by gravity before CO2 slugs catch up to generate foam in situ thereservoir.

© 2018 Published by Elsevier B.V. on behalf of The Korean Society of Industrial and EngineeringChemistry.

Contents lists available at ScienceDirect

Journal of Industrial and Engineering Chemistry

journal homepa ge: www.elsev ier .com/locate / j ie c

Introduction

CO2 injection has superior displacement efficiency wherevergas contacts oil. It recovers oil by mechanisms of extraction,dissolution, solubilization, and possibly some other phase behaviorchanges [1–3]. As of 2017 in United States, it produces 280,000barrels of oil per day accounting for approximately half of the oilproduced by enhanced oil recovery (EOR) methods and 6% of thetotal domestic oil production [4–7]. With novel technologies beingdeveloped to capture CO2 from industrial sites, CO2 injectionmethod is attracting more market interests. However due to thelarge density and viscosity contrast between CO2 and the crude oil,the sweep efficiency of CO2 flood is limited by gravity override andviscous fingering. The scenario could be worse if the reservoir ishighly heterogeneous or fractured where the fluids of highmobility tend to finger through a preferential path and leave alarge portion of the reservoir un-swept [2,8–16].

* Corresponding author.E-mail address: r.farajzadeh@tudelft.nl (R. Farajzadeh).

https://doi.org/10.1016/j.jiec.2018.12.0131226-086X/© 2018 Published by Elsevier B.V. on behalf of The Korean Society of Indus

The sweep efficiency of CO2 EOR can be improved by theinjection mode of water-alternating-gas (WAG) [17–22]. Ideally theinjected water can reduce the relative permeability of the CO2

phase and therefore delay the gas breakthrough. However, theimprovement can be limited because of phase segregation. A morepotent way to reduce the mobility of CO2 is to disperse the gas inaqueous phase with surfactants [8,23–30]. CO2 Foam in porousmedia is a dispersion of CO2 in aqueous phase such that theaqueous phase (wetting phase) is continuous and at least somepart of the CO2 (non-wetting phase) is made discontinuous by thinliquid films called lamella [31,8,32–35]. Depending on the reservoirtemperature and pressure, CO2 can be either gaseous-like or in thesupercritical state. Above the critical temperature (31.10 �C) andpressure (1071 psi), gaseous CO2 transitions into supercritical state.Conventional nomenclature refers to surfactant stabilized supercritical CO2 dispersion in water (C/W) as emulsion. Yet, forsimplicity, we do not differentiate the gaseous-CO2 foam and CO2-in-water (C/W) emulsion [36] and only use the term CO2 foamindicating that CO2 is the internal phase. The rationale is that thefundamental principles of interfacial phenomena and transportproperties as CO2 foam transitions into CO2 emulsion still hold the

trial and Engineering Chemistry.

Nomenclature

Csg Surfactant concentration in the gaseous phase (g/L)Csw Surfactant concentration in the aqueous phase (wt

%)epdry Foam model parameter that regulates how abrupt-

ly foam dries out at limiting water saturationepsurf Foam model parameter, the exponent in Fsurfactant

functionepoil Foam model parameter, the exponent in Foil

functionFoil Oil dependent function in foam modelFsurfactant Surfactant dependent function in foam modelFwater Water saturation dependent function in foam

modelfloil Foam model parameter that sets the maximum oil

saturation below which the oil has no impact onfoam

fmoil Foam model parameter that sets the minimum oilsaturation above which the oil kills the foamcompletely

FM Correction factor function for gas phase mobilityreduction by foam

fmmob Foam model parameter that sets the maximum gasmobility reduction

fmsurf Foam model parameter that sets the minimalsurfactant concentration above which foamstrength is no longer dependent on surfactantconcentration

kfrg Relative permeability to gas (foam)knfrg Relative permeability to gas (no foam)korg End-point relative permeability to gaskro Relative permeability to oilkorog End-point relative permeability to oil with respect

to gaskorow End-point relative permeability to oil with respect

to waterkrw Relative permeability to waterkorw End-point relative permeability to waterKsgw Surfactant partition coefficient between gaseous

phase and aqueous phaseng Corey exponent for gasnog Corey exponent for oil with respect to gasnow Corey exponent for oil with respect to waternw Corey exponent for waterPV Pore volumes injectedSg Gas saturationSgr Residual gas saturationSo Oil saturationSorg Residual oil saturation to gasSorw Residual oil saturation to waterSw Water saturationSwc Connate water saturationlfrg Mobility of the gas phase in presence of foam

lnfrg Mobility of the gas phase in absence of foam

134 Y. Zeng et al. / Journal of Industrial and Engineering Chemistry 72 (2019) 133–143

same. Admittedly, as the density increases, CO2 starts to solubilizecrude oil. Nevertheless, the compositional change of the oleicphase during CO2 flooding beyond minimal miscibility pressure(MMP) is beyond the scope of this paper. We mainly focus on thetransport of CO2 and surfactant in porous media.

Foam can decrease the mobility of CO2 and increase theapparent viscosity of injection fluids by blocking the continuous

gas path. Dispersed CO2 bubbles trapped in the foam structurehave a mobility that is orders of magnitude lower than continuousCO2 flow. Therefore foam assisted CO2 injection can effectivelyaddress the issue of poor mobility ratio and improve thevolumetric sweep efficiency.

Commonly used ionic surfactants can only be injected in waterbecause they are not soluble in the gas. Therefore the injectionmode for these surfactants is called surfactant-alternating-gas(SAG). Well-known ionic surfactants have been introduced indetail from literature [26,29,37–41]. Novel CO2 soluble surfactantsinclude the nonionic exthoxylated alcohols and switchableexthoxylated amines [27,40,42–49]. The hydrophobic part can beeither alkylphenol or branched/unbranched alkyl [45]. Thesesurfactants can be injected with CO2 phase. If surfactant is injectedwith CO2, we call it water-alternating-gas-plus-surfactant-in-gas(WAG + S) mode. WAG + S process has the potential to outperformSAG in different ways. Firstly WAG + S can improve the wellinjectivity [50] when surface facilities switch from CO2 injection towater injection. Secondly if surfactant is injected and transportedin the CO2 phase, it can foam with the water from secondaryrecovery, elongate the gas-water mixing zone and delay phasesegregation [27,51–53].

Because the surfactant can dissolve in both the CO2 and thewater, it is critical to understand how the partitioning of thesurfactant between the two phases can affect the foam transport inporous media. Partition coefficient Ksgw is a measure of the ratio insolubility of the surfactant in gaseous and aqueous phases. It isdefined as the ratio of the surfactant concentration in gaseousphase Csg to that in aqueous phase Csw at thermodynamicequilibrium as in Eq. (1). In some literature, surfactant partitioncoefficient is defined as the ratio of mass fraction rather than theconcentration. The two definitions of partition coefficient aredifferent by a factor that equals to the density ratio of the twophases, yet represent the same in nature.

Ksgw ¼ Csg

Cswð1Þ

The value of surfactant partition coefficient Ksgw is dependenton several factors [27,38,54] and can be experimentally measuredin the lab. Ren et al. [53] tested a series of different exthoxylatedalcohols and discovered that the value of the partition coefficientof these CO2 soluble surfactants varies in orders of magnitude withrespect to the reservoir conditions (from 0.02 to greater than 1). Ingeneral, the partition coefficient increases proportionally with thereservoir pressure, whereas decreases more dramatically with theincrease in reservoir temperature. Additionally, the Ksgw is verysensitive to surfactant formula and increases with decreasingethylene oxide (EO) groups. The number of propylene oxide (PO)group also plays a critical role in determining surfactant partitioncoefficient. Compared to EO groups, PO group is more hydrophobicand tends to increase the Ksgw. Unlike nonionic surfactants, thepartition coefficient of switchable surfactant is very sensitive to thereservoir pH. This is because the amine group can be protonated inthe aqueous phase and the protonation degree increases withdecreasing pH. This paper systematically simulates the transport ofsurfactant and foam in a 2-D homogeneous reservoir with differentKsgw values. In our previous paper [51], we established a 1-D 2-phase (water and CO2) home-made foam simulator and demon-strated that when surfactant is approximately equally partitioningbetween gaseous phase and aqueous phase, foam is in favor for oildisplacement in regard with apparent viscosity and foampropagation speed. In this paper, we extend the modeling to a2-D 3-phase (water, CO2, and oil) system with gravity at play. Thesimulation is done using Shell’s in-house Modular ReservoirSimulator (MoReS) [55]. We will briefly summarize the implicit-texture (IT) local-equilibrium (LE) foam model in Section “Implicit-

Fig. 1. Foam apparent viscosity mapp calculated as a function of foam quality(f g ¼ ug

ugþuw) and surfactant concentration Csw using parameters aforementioned

in Table 1.

Y. Zeng et al. / Journal of Industrial and Engineering Chemistry 72 (2019) 133–143 135

texture local-equilibrium foam model”, and review the three-phase relative permeability model in Section “Three-phase relativepermeability model”. We will introduce the 2-D homogeneousmodel reservoir created for simulation in Section “2-D Homoge-neous model reservoir for simulation”. In Section “CO2 displacingwater with different partition coefficients”, we will discuss theeffect of partition coefficient on foam by comparing the studiedscenarios in which continuous CO2 is injected with surfactants ofvaried partition coefficient into an aqueous reservoir to displacewater. In Section “Case study: WAG, SAG and WAG + S”, we willcompare the oil recovery efficiency between WAG, SAG andWAG + S modes and will show that CO2 soluble surfactants withproper partition coefficient can outperform conventional ionicsurfactants in terms of mobility control and oil recovery efficiencyby synchronizing the surfactant transport with gas phasepropagation in a 2-D homogeneous model system.

Numerical models

Implicit-texture local-equilibrium foam model

Foam can lower the gas phase mobility by orders of magnitude.Yet the water mobility is proven to remain the same at a givenwater saturation [56]. Apparent viscosity mapp is used as a measureof foam strength in porous media. It is defined as normalizedpressure gradient rp with respect to rock permeability k and totalsuperficial velocity (ug + uw) as shown in Eq. (2).

mapp ¼ � k � rpug þ uw

ð2Þ

The STARS version of the implicit-texture local-equilibriummodel [57,58] implemented here modifies the relative mobility tothe gas phase as shown in Eq. (3). The correction factor for the gasphase mobility reduction FM is inversely related to the product of aseries of dependence functions as shown in Eq. (4). The parameterfmmob sets a reference to the maximum foam strength. Thedependence functions Fsurfactant, Fwater, and Foil are all in the range of[0,1]. The surfactant concentration dependence function Fsurfactantand the water saturation dependence function Fwater are discussedin previous publications [59–62]. In Section “Case study: WAG, SAGand WAG + S” where the oil phase is present, function Foil isintroduced to account for the effect of oil on foam. Oil candestabilize foam in porous media by varied mechanisms[15,25,33,63,64]. In this simulation, the parameter floil representthe oil saturation below which oil does not affect foam strengthand Foil equals to 1, whereas the parameter fmoil represent the oilsaturation above which foam is completely killed and Foil equals to0. The parameter epoil regulates how Foil deceases as oil saturationSo increases floil to fmoil. We list the values assigned to theparameters aforementioned in Table 1 and show the correspond-ing foam apparent viscosity as a function of foam quality (gas

Table 1STARS foam model parameters.

Parameter Value

fmmob 500Fsurfactant fmsurf 0.2 wt%

epsurf 1Fwate fmdry 0.25

epdry 500Foil floil 0.1

fmoil 0.4epoil 1.5

fractional flow ug

ugþuw) and surfactant concentration Csw in Fig. 1.

lfrg ¼ lnf

rg � FM ð3Þ

FM ¼ 11 þ f mmob � Fsurf actant � Fwater � Foil

ð4Þ

Fsurf actant ¼Csw

f msurf

� �epsurf

f orCsw < f msurf

1 f orCsw � f msurf

8<: ð5Þ

Fwater ¼ 0:5 þ arctan½epdryðSw � f mdryÞ�p

ð6Þ

Foil ¼0 f or So > f moil

f moil � Sof moil � f loil

� �epoil

f or f loil < So < f moil

1 f or So < f loil

8>><>>:

ð7Þ

Three-phase relative permeability model

Corey model [65] is used to calculate the relative permeability

to water krw and gas knfrg (no foam). It is well-known that formationwettability plays a critical role in determining the relativepermeability curves [66–69]. Without introducing unnecessarycomplexities, we simply assume strict water-wet formationcondition in this paper. Therefore, the parameter krw is only a

function of water saturation Sw and knfrg is only a function of gassaturation Sg. In the 3-phase simulation, linear iso-perm is appliedto calculate the relative permeability to oil kro as shown in Fig. 2.

In the absence of gas, we use the water-oil two-phase Coreymodel to calculate kro based on the water/oil saturations (thickgreen line on the bottom); at connate water saturation (thick greenline parallel to the side Gas–Oil), we use the Gas–Oil Corey modelto calculate the kro at Swc. In the three phase region, the linear iso-perm model assumes that the saturations (Sw, Sg, So) on the sametie line (straight lines connecting the two thick green lines) give the

Fig. 2. Schematic of the linear iso-perm relative permeability to oil in 3-phase region.

136 Y. Zeng et al. / Journal of Industrial and Engineering Chemistry 72 (2019) 133–143

same kro value. The three-phase relative permeability parametersare listed in Table 2.

2-D Homogeneous model reservoir for simulation

A 2-D homogeneous model reservoir with a permeability of 1Darcy is created for the simulations in Section “CO2 displacingwater with different partition coefficients” and “Case study: WAG,SAG and WAG + S”. The model reservoir is 2000 ft in length and200 ft in thickness. Initially, the reservoir condition is set at 100 barin pressure and 100 �C in temperature. In all cases studied, fluidsare injected at 1 ft/day interstitial velocity. The Peclet numbers forthe surfactant dispersion in both phases (CO2 and water) are equalto 50 for field-level simulation [2].

In Section “CO2 displacing water with different partitioncoefficients”, the reservoir is initially 100% saturated with water.CO2 is injected continuously with surfactant of different Ksgw valuesto displace the water as shown in Table 3. The injection surfactantconcentration is 2.5 g/L CO2 at reservoir conditions,1.25 times of thefmsurf value. In this Section, we will focus on the effect of surfactantpartitioning between CO2 and aqueous phases on the surfactant

Table 2Three-phase relative permeability parameters to water, oil, and gas (no foam).

Water–oil relative permeability parameters

Parameter Symbol Value

Connate water saturation Swc 0.10Residual oil saturation to water Sorw 0.40End-point relative permeability to water korw 0.22

End-point relative permeability to oil korow 1.0

Corey exponent for water nw 4.0Corey exponent for oil with respect to water now 2.0

Oil–gas relative permeability parametersParameter Symbol ValueResidual gas saturation Sgr 0.05Residual oil saturation to gas Sorg 0.01End-point relative permeability to gas korg

1.00

End-point relative permeability to oil korog korowCorey exponent for gas ng 1.7Corey exponent for oil with respect to gas nog 4.0

transport and foam propagation, and investigate the optimalpartition coefficient that maximizes the sweep efficiency of CO2.

In Section “Case study: WAG, SAG and WAG + S”, we add onemore degree of complexity by introducing black oil into the modelsystem. The hypothetical black oil has an API gravity of 45� and aviscosity of 0.4 cP under the reservoir condition (100 �C and100 bar). The reservoir is initially at an oil saturation Soi= 1 � Swcð Þand the reservoir is water flooded for 1 PV before enhanced oilrecovery techniques are applied. After the water flooding, theaverage oil saturation is reduced to 0.4. WAG, SAG, and WAG + S areapplied to the water flooded reservoir respectively in Case A, B, andC. In this Section, we will focus on the comparison between theseEOR applications.

Results and discussion

In this Section, we will first show the results of continuous CO2

injection with CO2 soluble surfactant to displace water and thencompare the various EOR techniques, highlighting the benefit ofinjecting surfactant with CO2 and the critical role of Ksgw.

CO2 displacing water with different partition coefficients

In Case I, CO2 with a CO2 soluble surfactant of small partitioncoefficient (Ksgw = 0.01) is injected to displace water. The injectedsurfactant concentration is 2.5 g/L at reservoir condition. Fig. 3displays the snapshots of the saturation and the surfactantconcentration profile at different dimensionless times, the totalpore volumes (PV) of CO2 injected. It appears that the surfactanttransport is severely retarded with respect to the gas propagation.After 1PV of injection, the surfactant only penetrates a fairly smallportion of the reservoir. The segregation between the surfactant andthe CO2 is because of the extremely high surfactant affinity to water.Upon water contact, most of the surfactant quickly partitions intothe aqueous phase. Therefore the surfactant accumulates in the nearwell-bore region and the surfactant concentrationin thewater Csw ispredominately higher than that in the gas phase Csg. Once thesurfactant is stripped off to the water, the CO2 mobility control islost. Consequently, the gas overrides the upper layer of the reservoirand streaks through early and the sweep of CO2 is poor.

In Case II, the surfactant partition coefficient is set to 1, whichmeans that the surfactant has equal affinity to both the CO2 and the

Table 3Characteristic values for surfactant partition coefficient Ksgw for the case studies inboth Section “CO2 displacing water with different partition coefficients” andSection “Case study: WAG, SAG and WAG + S”.

Surfactant partitioncoefficient

Characteristicvalue

Remark

Small Ksgw 0.01 Strong affinity to waterUnity Ksgw 1.00 Equal affinity to water and

CO2

Large Ksgw 50.0 Strong affinity to CO2

Fig. 3. Case I:Continuous CO2 (with dissolved surfactant) injection to displace water witoverrides the reservoir and prematurely breaks through; (B) Surfactant concentration

transport is retarded.

Fig. 4. Case II: Continuous CO2 (with dissolved surfactant) injection to displace water wsweep of CO2 is greatly improved; (B) Surfactant concentration profile indicating that

Y. Zeng et al. / Journal of Industrial and Engineering Chemistry 72 (2019) 133–143 137

water. In contact with water, the surfactant concentration in thegas phase Csgwill be equal to that in the water Csw. Fig. 4 shows thesnapshots of the surfactant distribution along with the phasesaturation profile. In this case, the surfactant transport issynchronized with the gas propagation. The surfactant createsfoam with the water residing in the reservoir wherever the gassweeps. The foamed gas can effectively mitigate the effect ofgravity. Therefore, the sweep efficiency of CO2 is greatly improved.After 1 PV of CO2 injection, most of the reservoir has been swept.

In case III, the surfactant partition coefficient is increased to 50,which means that the surfactant prefers to stay with the CO2 even

h small partition coefficient Ksgw = 0.01 (A) Saturation profile indicating that the gasprofile indicating that the surfactant is highly concentrated near the well and the

ith unity partition coefficient Ksgw = 1.00 (A) Saturation profile indicating that thethe surfactant transport is synchronized with the CO2 propagation.

138 Y. Zeng et al. / Journal of Industrial and Engineering Chemistry 72 (2019) 133–143

if it is equilibrated with water. At equilibrium, the surfactantconcentration in the CO2 Csg will be a lot higher than that in thewater Csw. From Fsurfactant function, it is clear that foam strength isdirectly related to the Csw. Therefore, the highly diluted surfactantconcentration in the aqueous phase results in insufficient gasmobility reduction and the foam is not strong enough to fightagainst the gravity. Consequently, only the upper layer of thereservoir is swept by CO2. Unlike Case I in which the CO2 loses itsmobility control completely, in Case III, the poor sweep efficiencyresults from the lack of surfactant in the aqueous phase as shownin Fig. 5.

From the case studies above, we conclude that the foamtransport is highly dependent on the surfactant partition coeffi-cient. Fig. 6 compares the water recovery efficiency with respect todifferent partition coefficient. Before the gas breakthrough, theefficiency of the water recovery is proportional to the volume ofCO2 injected. Therefore the three water recovery curves overlap inthe beginning. In case of unity (Ksgw = 1) and large (Ksgw = 50)partition coefficients, most of the surfactant is carried by the CO2.Therefore, after the CO2 breaks through, the water recoveryefficiency tends to level off and reach plateau. However, in the caseof small partition coefficient (Ksgw = 0.01), the water recoveryefficiency keeps increasing after the CO2 breakthrough. This is dueto the propagation of retarded surfactant bank. Because of the highaffinity of the surfactant to the aqueous phase, water strips off thesurfactant from the CO2 soon after injection. Therefore, thesurfactant front is left far behind the gas front. When the CO2

reaches the production well at the breakthrough time, surfactant isstill propagating in the reservoir. Therefore, more foam is beingcreated and the water recovery efficiency keeps increasing.However, it takes much more volumes of CO2 to reach the samelevel of water recovery efficiency compared to the case of unitypartition coefficient.

In summary, unity partition coefficient is superior to either toolarge or too small partition coefficient in terms of mobility controland maximizing sweep efficiency. When the surfactant partitioncoefficient is too small, the surfactant will be highly concentratednear the well-bore, and the gas breaks through early. Too largepartition coefficient result in weak foam, and the foam strengthmight not suffice to fight against gravity effect.

Fig. 5. Case III: Continuous CO2 (with dissolved surfactant) injection to displace water wfoam strength is insufficient to keep CO2 from overriding the reservoir; (B) Surfactant conis highly diluted.

Case study: WAG, SAG and WAG + S

In this Section, we will compare three EOR techniques (Case A:WAG, Case B: SAG, and Case C: WAG + S) to demonstrate theadvantage of using conventional foam and foam with CO2 solublesurfactant for residual oil recovery. In all cases, the reservoir iswater flooded to the remaining oil saturation of 0.4. In Case A,WAG, water slugs and gas slugs are alternatively injected into thereservoir. Each cycle consists of 0.2 PV of gas and 0.05 PV of water.In other words, the average gas fractional flow is 0.8. In Case B, SAG,0.5 wt% of traditional surfactant (only soluble in water) is added tothe aqueous phase to generate foam in situ. All the other operationconditions are kept exactly the same as Case A. In Case C, WAG + S,the same amount of CO2 soluble surfactant is injected with CO2

instead of water. The only difference between Case B and C is thatin Case C, the surfactant is injected with the gas phase andpartitions to the aqueous phase to foam with a partition coefficientof unity (Ksgw = 1.0). The foam model parameters in Case B and C arekept exactly the same, meaning that the surfactants have the samefoaming capability.

Fig. 7 displays the 3-phase saturation profile snapshots for theWAG process. The gas only floods the upper part of the reservoirand breaks through during the first gas slug injection. Thesaturation profile does not change much after the first cycle isinjected. Due to the density and viscosity contrast, the CO2 pushesthe oil down. An oil band is formed below the gas path way,however, is hardly produced due to the poor gas mobility control.

Foam can effectively mitigate the gravity effect. Fig. 8 shows theresults of Case B in which 0.5 wt% surfactant (only soluble in water)is added to the aqueous phase. Gas mobility is significantly reducedby the generation of foam inside the porous medium. Gas (depictedin red) partially penetrates the lower part of the reservoir and anoil bank in green is formed in front of the foam. The oil is slowlyproduced from the production well on the right.

However, comparing the 3-phase saturation profile and thesurfactant concentration profile in Fig. 8, it is noticed thatsignificant amount of surfactant drains down before the gas slugscatches up and is wasted during the water injection period. The gasslug overrides the reservoir; however, the surfactant slug under-rides the reservoir. A better alternative to inject the surfactant is

ith large partition coefficient of Ksgw = 50.0 (A) Saturation profile indicating that thecentration profile indicating that the surfactant concentration in the aqueous phase

Fig. 6. Water recovery efficiency comparison between cases of varied surfactant partition coefficients. Unity partition coefficient is superior to either too small or too largepartition coefficient in terms of water recovery efficiency.

Fig. 7. Case A: Water-alternating-gas. The 3-phase saturation profile indicating that WAG mode has very limited improvement in oil recovery compared to waterflooding.

Y. Zeng et al. / Journal of Industrial and Engineering Chemistry 72 (2019) 133–143 139

needed to synchronize the gas propagation and the surfactanttransport such that the surfactant is available to make foam wherethe gas sweeps.

Fig. 9 displays the results when the same amount of surfactantis injected with the CO2 phase instead of the water. The partitioncoefficient between the water and the CO2 in this case is unity(Ksgw = 1.0). With the favorable partition coefficient, it is shown thatthe sweep is further improved from the SAG process. The gas phasecarries the surfactant and dissolves the right amount to water togenerate foam. It is notable that the gas penetrates into an evenlarger area in lower part of the reservoir. In addition, the oil bankaccumulated is larger compared to the SAG process. All theseimprovement is because of the synchronization of the gaspropagation and the surfactant transport.

However, it is important to point out that the recoveryefficiency of WAG + S is highly dependent on the partitioncoefficient. As discussed in Section “CO2 displacing water withdifferent partition coefficients”, too small partition coefficientresults in surfactant transport retardation whereas too largepartition coefficient results in insufficient foam robustness. Asshown in Fig. 10(A), partition coefficient of either 0.01 or 50.0 leadsto poorer oil recovery efficiency compared to SAG process. Thesensitivity of surfactant partition coefficient on oil recovery isdisplayed in Fig. 10(B). The shape of the oil recovery efficiencycurve, exhibiting a local maximum, is a result of the competingmechanisms between faster surfactant transport (large Ksgw) and

higher foam strength (small Ksgw). The surfactant needs to betransported with the CO2 whereas adequate amount of thesurfactant needs to be dissolved in the water to make the foamstrong. The optimized surfactant partition coefficient for WAG + Sprocess is unity for the case considered here.

To further investigate the WAG + S process, a series of sensitivityanalysis are conducted with respect to a number of criticalreservoir properties such as crude oil viscosity, Peclet number/dispersivity, reservoir thickness, and well spacing (distancebetween the injection well and the production well). In all thesensitivity analysis, the reservoir is water-flooded for 1 PV and thenthe same WAG + S process as described above with a unity partitioncoefficient (Ksgw = 1.0) was applied. The overall oil recoveryefficiency (crude oil percentage recovered based on the originaloil in place (OOIP)) is plotted as a function of different variables asshown in Fig. 11.

Fig. 11(A) shows that the overall oil recovery efficiency is astrong function of the crude oil viscosity. As the crude oil viscosityincreases, the efficiency of WAG + S process decreases linearly as afunction of the logarithm of the crude oil viscosity using the sameset of foam parameters. To generalize the observation, we define anew dimensionless variable called the oil/foamed gas viscosityratio as shown in Eq. (8). The numerator is the crude oil viscositymoil whereas the denominator is the gas viscosity mgas times theparameter fmmob. The denominator is a characteristic of themaximum foam strength (gas mobility reduction) that can be

Fig. 8. Case B: Surfactant-alternating-gas. (A) 3-phase saturation profile indicating that the gas mobility is reduced by foam in presence of surfactant and the sweep efficiencyis greatly improved from WAG injection. (B) Surfactant concentration profile indicating that significant surfactant drains by gravity before the CO2 slug catches up. (Forinterpretation of the references to colour in the text, the reader is referred to the web version of this article.)

Fig. 9. Case C: Water-alternating-gas-plus-surfactant-in-gas. (A) 3-phase saturation profile indicating that a larger oil bank is formed and the oil recovery is further improvedfrom SAG mode. (B) Surfactant concentration profile indicating that the surfactant transport is synchronized with the gas phase propagation.

140 Y. Zeng et al. / Journal of Industrial and Engineering Chemistry 72 (2019) 133–143

expected. It is a simple measure of the relative viscosity of thecrude oil and foam. It is concluded that stronger foam will berequired as the crude oil viscosity increases in order to achieve agiven oil recovery efficiency. Yet, it is worthwhile to mention thatthe crude oils of different viscosity are likely to have differentfoam-oil interactions. Therefore the parameters in the oilsaturation dependent function Foil are likely to vary. For simplicity,such effect is not included in this analysis.

Oil=Foamed Gas Viscosity Ratio ¼ moil

mgas � f mmobð8Þ

Fig. 11(B) shows the sensitivity analysis with respect to thedispersion. The overall oil recovery efficiency is plotted as afunction of both the Peclet number and the dispersivity of thesurfactant in the reservoir. It is well discussed in the literature [2]that the actual dispersivity of components increases with thecharacteristic length of the system. The reason behind is that thegeological formations have heterogeneities of all length scales andthus mixing at all length scales. However, in this sensitivityanalysis, it is found that the efficiency of WAG + S process is hardlyaffected by a wide change in the dispersivity values. In both labscale and reservoir scale dispersion, the overall oil recovery

Fig. 10. (A) Oil recovery efficiency comparison between WAG, SAG, and WAG + S with varied partition coefficient. (b) Sensitivity of surfactant partition coefficient on WAG + Soil recovery efficiency.

Fig.11. Sensitivity analysis: (A) Overall oil recovery efficiency as function of crude oil viscosity and oil/foamed gas viscosity ratio. (B) Overall oil recovery efficiency as functionof Peclet number and dispersity of the surfactant. (C) Overall oil recovery efficiency as function of reservoir thickness. (D) Overall oil recovery efficiency as function of wellspacing (distance between injection and production wells).

Y. Zeng et al. / Journal of Industrial and Engineering Chemistry 72 (2019) 133–143 141

efficiency is almost invariant as a function of the Peclet numberand the dispersivity length scale. Yet, it is worthwhile to point outthat the Peclet number can play a more pronounced role with smallpartition coefficient when surfactant is highly concentrated nearthe wellbore region. In that case, small Peclet number helps the

transport/propagation of surfactant species and can potentiallyimprove the sweep efficiency.

Fig.11(C) and (D) show that the WAG + S foam process is also notvery sensitive to the change in reservoir thickness and wellspacing. As long as strong foam is generated in the presence of

142 Y. Zeng et al. / Journal of Industrial and Engineering Chemistry 72 (2019) 133–143

crude oil, the process is effective across a wide range of reservoirdimensions. Admittedly, all our simulation is done at injection rateconstraint. Sparse well spacing requires higher foam injectivitywhich might not be achievable at reservoir conditions.

Conclusions

We simulated the transport of CO2 soluble surfactant in porousmedia and probed the effect surfactant partition coefficient on foamEOR. We firstly injected CO2 with surfactant of different Ksgw valuesto a 100% water saturated reservoir. It is found out that the sweep ofCO2 is highly dependent on Ksgw values. Similar to our previouspublication in 1-D system, when the surfactant has equal affinity toboth the water and the gas, the sweep efficiency is optimized. A 2-Dthick homogeneous reservoir with residual light oil was then set upfor numerical simulation to probe the use of CO2 soluble surfactantfor foam EOR. In our case, WAG process has marginally betterrecovery efficiency compared to water flooding because of the lowdensity and viscosity of the injected CO2 phase. SAG processimproves recovery efficiency by foaming the gas with surfactantsolution and therefore providing mobility control. Yet, in the case ofSAG injection, significant surfactant is wasted when the surfactantdrains with the water before the following gas slug catches up tofoam. A better alternative to SAG is WAG + S in which the surfactantis dissolved in the CO2 phase and injected with gas instead of water.WAG + S can synchronize the transport of surfactant with the gasphase propagation when the partition coefficient is around 1. Aseries of sensitivity analysis were conducted to understand theefficiency and robustness of the WAG+S process. It is found that theoil recovery efficiency is a strong function of the crude oil viscosityor the oil/foamed gas viscosity ratio. However, the efficiency of theprocess is not severely impacted with changes in dispersion,reservoir thickness, and/or well spacing in our investigated range.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgement

We acknowledge the financial support from Shell GlobalSolutions International (Rijswijk, Netherlands), Rice UniversityConsortium in Processes in Porous Media (Houston, TX).

References

[1] V. Alvarado, E. Manrique, Energies 3 (9) (2010) 1529.[2] L. Lake, R. Johns, W.R. Rossen, G. Pope, Soc. Pet. Eng. (2014).[3] M. Orr. Franklin, Theory of Gas Injection Processes, Tie-Line Publications,

Copenhagen, Denmark, 2017.[4] V.A. Kuuskraa, M.L. Godec, P. Dipietro, Energy Procedia 37 (2013) 6854.[5] D. Hussain, D.A. Dzombak, P. Jaramillo, G.V. Lowry, Int. J. Greenh. Gas Control

16 (2013) 129.[6] P. Jaramillo, W.M. Griffin, S.T. McCoy, Environ. Sci. Technol. 43 (21) (2009) 8027.[7] Y. Zeng, A Multiscale Study of Foam: The Phase Behavior, Transport, and

Rheology of Foam in Porous Media. Ph.D. Thesis, Rice University, 6100 Main St,Houston, TX, 2017 77005.

[8] A.R. Kovscek, C.J. Radke, Foams Fundamentals and Applications in thePetroleum Industry, vol. 242, American Chemical Society, 1994.

[9] S.I. Kam, W. Rossen, SPE J. 8 (04) (2003) 417.[10] C.A. Conn, K. Ma, G.J. Hirasaki, S.L. Biswal, Lab Chip 14 (20) (2014) 3968.[11] R.F. Li, W. Yan, S. Liu, G. Hirasaki, C.A. Miller, SPE J. 15 (04) (2010) 928.[12] M.A. Fernø, L.P. Hauge, A. Uno Rognmo, J. Gauteplass, A. Graue, Geophys. Res.

Lett. 42 (18) (2015) 7414.[13] M.A. Fernø, Ø. Eide, M. Steinsbø, S.A.W. Langlo, A. Christophersen, A. Skibenes,

T. Ydstebø, A. Graue, J. Pet. Sci. Eng. 135 (2015) 442.[14] R. Singh, K. Mohanty, Fuel 197 (2017) 58.[15] S. Xiao, Y. Zeng, E.D. Vavra, P. He, M.C. Puerto, G.J. Hirasaki, S.L. Biswal,

Langmuir 34 (3) (2017) 739.[16] P. Dong, M. Puerto, G. Jian, K. Ma, K. Mateen, G. Ren, G. Bourdarot, D. Morel, M.

Bourrel, S.L. Biswal, et al., SPE J. (2018).[17] M.M. Kulkarni, D.N. Rao, J. Pet. Sci. Eng. 48 (1–2) (2005) 1.

[18] M.F. Sedaralit, N.B. Darman, R.D. Tewari, R.Z. Kamarul Bahrim, P. Abdul Hamid,S.R. Mohd Shafian, A.A. Abdul Manap, Enhancing the Efficiency of ImmiscibleWater Alternating Gas (WAG) Injection in a Matured, High Temperature andHigh CO2 Solution Gas Reservoir - A Laboratory Study, SPE Enhanced OilRecovery Conference, 2–4 July, Kuala Lumpur, Malaysia, 2013 (165303-MS).

[19] M. Sohrabi, D.H. Tehrani, A. Danesh, G.D. Henderson, SPE J. 9 (03) (2004) 290.[20] J.D. Rogers, R.B. Grigg, SPE Reservior Eval. Eng. 4 (05) (2001) 375.[21] L. Han, Y. Gu, Energy Fuels 28 (11) (2014) 6811.[22] H. Lei, S. Yang, L. Zu, Z. Wang, Y. Li, Energy Fuels 30 (11) (2016) 8922.[23] T. Blaker, M.G. Aarra, A. Skauge, L. Rasmussen, H.K. Celius, H.A. Martinsen, F.

Vassenden, SPE Reservior Eval. Eng. 5 (04) (2002) 317.[24] A.H. Falls, J.B. Lawson, G.J. Hirasaki, J. Pet. Technol. 40 (01) (1988) 95.[25] K. Mannhardt, I. Svorstøl, J. Pet. Sci. Eng. 23 (3–4) (1999) 189.[26] L. Cui, K. Ma, M. Puerto, A.A. Abdala, I. Tanakov, L.J. Lu, Y. Chen, A. Elhag, K.P.

Johnston, S.L. Biswal, et al., SPE J. (2016).[27] G. Ren, Q.P. Nguyen, Pet. Sci. (2017).[28] S. Kahrobaei, S. Vincent-Bonnieu, R. Farajzadeh, Sci. Rep. 7 (1) (2017).[29] G.J. Hirasaki, C.A. Miller, M. Puerto, et al., Recent Advances in Surfactant EOR,

SPE Annual Technical Conference and Exhibition, 21–24 September, Denver,Colorado, USA, 2008 (SPE-115386-MS).

[30] G.J. Hirasaki, C.A. Miller, R. Szafranski, J.B. Lawson, N. Akiya, Surfactant/FoamProcess for Aquifer Remediation, International Symposium on OilfieldChemistry, Houston, Texas, 1997 18–21 February, (SPE-37257-MS).

[31] S. Alzobaidi, C. Da, V. Tran, M. Prodanovi�c, K.P. Johnston, J. Colloid Interface Sci.488 (2017) 79.

[32] A. Al Sumaiti, A.R. Shaik, E.S. Mathew, W. Al Ameri, Energy Fuels 31 (5) (2017)4637.

[33] R. Farajzadeh, A. Andrianov, R. Krastev, G.J. Hirasaki, W.R. Rossen, Adv. ColloidInterface Sci. 183–184 (0) (2012) 1.

[34] Y. Zeng, R. Farajzadeh, A.A. Eftekhari, S. Vincent-Bonnieu, A. Muthuswamy, W.R. Rossen, G.J. Hirasaki, S.L. Biswal, Langmuir (2016).

[35] Y. Zeng, R.Z.K. Bahrim, S.V. Bonnieu, J. Groenenboom, S.R.M. Shafian, A.A.A.Manap, R.D. Tewari, S.L. Biswal, Offshore Technology Conference Asia, KualaLumpur, Malaysia, 20-23 March, The Dependence of Methane Foam Transporton Rock Permeabilities and Foam Simulation on Fluid Diversion inHeterogeneous Model Reservoir, 20–232018 (OTC-28229-MS).

[36] S. Kahrobaei, K. Li, S. Vincent-Bonnieu, R. Farajzadeh, AIChE J. 64 (2) (2018) 758.[37] K. Mannhardt, J.J. Novosad, K.N.N. Jha, J. Can. Pet. Technol. 33 (02) (1994).[38] Y. Chen, A.S. Elhag, L. Cui, A.J. Worthen, P.P. Reddy, J.A. Noguera, A.M. Ou, K. Ma,

M. Puerto, G.J. Hirasaki, et al., Ind. Eng. Chem. Res. 54 (16) (2015) 4252.[39] M. Chorro, N. Kamenka, B. Faucompre, S. Partyka, M. Lindheimer, R. Zana,

Colloids Surf. Physicochem. Eng. Aspects 110 (3) (1996) 249.[40] G. Jian, M.C. Puerto, A. Wehowsky, P. Dong, K.P. Johnston, G.J. Hirasaki, S.

Biswal, Langmuir 32 (40) (2016) 10244.[41] G. Jian, M. Puerto, A. Wehowsky, C. Miller, G.J. Hirasaki, S.L. Biswal, J. Colloid

Interface Sci. 513 (2018) 684, doi:http://dx.doi.org/10.1016/j.jcis.2017.11.041.[42] Y. Chen, A.S. Elhag, B.M. Poon, L. Cui, K. Ma, S.Y. Liao, P.P. Reddy, A.J. Worthen, G.

J. Hirasaki, Q.P. Nguyen, et al., SPE J. 19 (02) (2014) 249.[43] A.S. Elhag,Y. Chen,H. Chen, P.P. Reddy, L. Cui, A.J. Worthen, K. Ma,G.J. Hirasaki, Q.P.

Nguyen, S.L. Biswal, et al., Switchable AmineSurfactants for Stable C/Brine Foamsin High Temperature, High Salinity Reservoirs SPE Improved Oil RecoverySymposium, Tulsa, Oklahoma, USA 12-16 April, 2014 (SPE -169041-MS).

[44] L. Cui, K. Ma, A.A. Abdala, L.J. Lu, I. Tanakov, S.L. Biswal, G.J. Hirasaki, Adsorptionof a Switchable Cationic Surfactant on Natural Carbonate Minerals; SPEImproved Oil Recovery Symposium, Tulsa, Oklahoma, USA, 12-16 April, 2014(SPE -169040-MS).

[45] D. Xing, B. Wei, W.J. McLendon, R.M. Enick, S. McNulty, K. Trickett, A.Mohamed, S. Cummings, J. Eastoe, S. Rogers, et al., SPE J. 17 (04) (2012) 1172.

[46] Y. Zhang, L. Zhang, Y. Wang, M. Wang, Y. Wang, S. Ren, Chem. Eng. Res. Des. 94(2015) 624.

[47] Y. Wang, Y. Zhang, Y. Liu, L. Zhang, S. Ren, J. Lu, X. Wang, N. Fan, J. Pet. Sci. Eng.154 (2017) 234.

[48] Y. Zhang, Y. Wang, F. Xue, Y. Wang, B. Ren, L. Zhang, S. Ren, J. Pet. Sci. Eng. 133(2015) 838, doi:http://dx.doi.org/10.1016/j.petrol.2015.04.003.

[49] S.S. Adkins, X. Chen, I. Chan, E. Torino, Q.P. Nguyen, A.W. Sanders, K.P. Johnston,Langmuir 26 (8) (2010).

[50] D. Shan, W.R. Rossen, Optimal Injection Strategies for Foam IOR; SPE/DOEImproved Oil Recovery Symposium, Tulsa, Oklahoma, 13-17 April, 2002 (SPE-75180-MS).

[51] Y. Zeng, K. Ma, R. Farajzadeh, M. Puerto, S.L. Biswal, G.J. Hirasaki, Transp. PorousMedia (2016).

[52] E. Ashoori, T. van der Heijden, W.R. Rossen, Fractional Flow Theory of FoamDisplacements With Oil, SPE International Symposium on Oilfield ChemistryThe Woodlands. Texas 20-22 April, 2009 (SPE -121579-MS).

[53] G. Ren, A.W. Sanders, Q.P. Nguyen, J. Supercrit. Fluids 91 (2014) 77.[54] C.T. Lee, P.A. Psathas, K.P. Johnston, J. deGrazia, T.W. Randolph, Langmuir 15

(20) (1999) 6781.[55] J.M.M. Regtien, G.J.A. Por, M.T. van Stiphout, F.F. van der Vlugt, Interactive

Reservoir Simulation, SPE Reservoir Simulation Symposium San Antonio,Texas 12-15 February, 1995 (SPE -29146-MS).

[56] A.A. Eftekhari, R. Farajzadeh, Sci. Rep. 7 (2017) 43870.[57] K. Ma, G. Ren, K. Mateen, D. Morel, P. Cordelier, SPE J. 20 (03) (2015) 453.[58] M. Lotfollahi, R. Farajzadeh, M. Delshad, A. Varavei, W.R. Rossen, J. Nat. Gas Sci.

Eng. 31 (2016) 184.[59] K. Ma, J.L. Lopez-Salinas, M.C. Puerto, C.A. Miller, S.L. Biswal, G. Hirasaki, Energy

Fuels 27 (5) (2013) 2363.

Y. Zeng et al. / Journal of Industrial and Engineering Chemistry 72 (2019) 133–143 143

[60] Y. Zeng, A. Muthuswamy, K. Ma, L. Wang, R. Farajzadeh, M. Puerto, S. Vincent-Bonnieu, A.A. Eftekhari, Y. Wang, C. Da, et al., Ind. Eng. Chem. Res. 55 (28)(2016) 7819.

[61] O.G. Apaydin, A.R. Kovscek, Transient Foam Flow in Homogeneous PorousMedia: Surfactant Concentration and Capillary End Effects, SPE/DOE ImprovedOil Recovery Symposium Tulsa, Oklahoma 3-5 April, 2000 (SPE -59286-MS).

[62] B. Géraud, S.A. Jones, I. Cantat, B. Dollet, Y. Méheust, Water Resour. Res. 52 (2)(2016) 773.

[63] M. Amirmoshiri, Y. Zeng, Z. Chen, P.M. Singer, M.C. Puerto, H. Grier, R.Z.K.Bahrim, S. Vincent-Bonnieu, R. Farajzadeh, S.L. Biswal, et al., Energy Fuels 32(11) (2018) 11177.

[64] E.D. Vavra, Y. Zeng, S. Xiao, G.J. Hirasaki, S.L. Biswal, J. Visualized Exp. (131)(2018).

[65] A.T. Corey, Prod. Mon. (19) (1954) 38.[66] W. Abdallah, J.S. Buckley, A. Carnegie, J. Edwards, B. Herold, E. Fordham, A.

Graue, T. Habashy, N. Seleznev, C. Signer, et al., Technology 38 (1125–1144)(1986) 268.

[67] G. Hirasaki, Wettability Fundamentals and Surface Forces, vol. 6(1991) SPEFormation Evaluation Issue 02, June (SPE-17367-PA).

[68] J. Song, Y. Zeng, L. Wang, X. Duan, M. Puerto, W.G. Chapman, S.L. Biswal, G.J.Hirasaki, J. Colloid Interface Sci. 506 (2017) 169.

[69] B.M. Lescure, E.L. Claridge, CO2 Foam Flooding Performance vs. RockWettability, SPE Annual Technical Conference and Exhibition, New Orleans,Louisiana, 5-8 October, 1986 (SPE-15445-MS).